TREE NEWS update: Microsoft released Microsoft-Decision-1 on October 12, a new decision model post-trained on Alibaba’s Qwen3.5-9B. Microsoft said the model outperforms popular decision models including LLM and Jev 1.13.0 in latency and accuracy on structured decision tasks. The Xbox research team has used it to process over 10,000 pieces of player feedback, and CEO Nadella said Microsoft is testing it across all areas, including incident response, quality control and scientific discovery.
Microsoft Launches Decision Model Built on Alibaba’s Qwen
The notable signal is procurement, not performance: a US hyperscaler shipping a product built on a Chinese open-weight base model suggests in-house frontier training is not the only viable path for narrow, task-specific systems. That matters for the Qwen ecosystem's credibility in enterprise pipelines and for Western labs positioning proprietary models as the default. Whether this reflects a deliberate multi-vendor sourcing strategy or a one-off experiment is the open question, and how Microsoft frames the dependency in future releases will indicate which.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.